Metraipolis › archive › dc_df5a4e5c3a95
REWIND: Uncertainty-Guided Trajectory Repair for Test-Time Reasoning — arXiv:2610.19561
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.19561
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T20:02:42.220Z
- SHA-256 of the archived text
c7b0e23163ecb91fde1eceab8a4a9050306817c5188573a75c142b70e62c845b- Stable URL of this page
https://metraipolis.com/documento/dc_df5a4e5c3a95
How to cite
REWIND: Uncertainty-Guided Trajectory Repair for Test-Time Reasoning — arXiv:2610.19561 — arXiv (en), https://arxiv.org/abs/2610.19561 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_df5a4e5c3a95 (Metraipolis archive dc_df5a4e5c3a95, sha256 c7b0e23163ecb91fde1eceab8a4a9050306817c5188573a75c142b70e62c845b)
BibTeX
@misc{dcdf5a4e5c3a95,
title = {REWIND: Uncertainty-Guided Trajectory Repair for Test-Time Reasoning — arXiv:2610.19561},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.19561},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_df5a4e5c3a95; sha256 c7b0e23163ecb91fde1eceab8a4a9050306817c5188573a75c142b70e62c845b},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19561. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
An early mistake in the generated reasoning can constrain all subsequent reasoning. Repeated-generation methods, like MCMC-based approaches, attempt to improve output quality by sampling multiple generations, but their effectiveness can be limited by randomly selecting where to regenerate. On MATH500 with Qwen2.5-Math-7B, we find that incorrect responses have substantially higher mean token entropy than correct responses, with an average within-question relative increase of approximately 112%, suggesting that entropy can signal reasoning errors. Motivated by this finding, we introduce REWIND, a training-free controller that uses entropy dynamics to propose targeted suffix regenerations. Across three model families and five benchmarks, REWIND outperforms all baselines, with gains of 0.51-10.00 percentage points over the strongest baseline. These results demonstrate the effectiveness of targeted, uncertainty-guided trajectory repair across models and tasks. Our code is available at this https URL . Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.19561 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-15T19:35:48.348Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).